Merge pull request #36 from BillyMich/sampling

Sampling merge knn version
BillyMich · Nov 30, 2023 · 27437e6 · 27437e6
2 parents 778dbc1 + b3baee2
commit 27437e6
Show file tree

Hide file tree

Showing 10 changed files with 108 additions and 36 deletions.
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
diff --git a/datasets/00020000-1.bin b/datasets/00020000-1.bin
diff --git a/datasets/00050000-3.bin b/datasets/00050000-3.bin
diff --git a/include/knn_improvments.h b/include/knn_improvments.h
@@ -3,15 +3,16 @@
 
 #include "../include/knn.h"
 
-void knn_improved_algorithm(Graph** graph, int K, String distance_function);
+void knn_improved_algorithm(Graph** graph, int K, String distance_function, double p);
 
 void changeNeighbors(Graph** graph);
 
-void localJoin(Node** node, String distance_function);
+void localJoin(Node** node, String distance_function, int pK);
 
 int incrementalSearch(NodeNeighborsLinkedList* neighbor1, NodeNeighborsLinkedList* neighbor2);
 
-//sampling
+NodeNeighborsLinkedList* sampling(NodeNeighborsLinkedList* neighbors, int pK);
+
 //earlyTermination
 
 #endif
diff --git a/main/Makefile b/main/Makefile
@@ -18,9 +18,9 @@ OBJS = main.o $(MODULES)/Graph.o $(MODULES)/Node.o $(MODULES)/Neighbors.o $(MODU
 EXEC = program
 
 # Args for examples
-ARGSEUCL = $(DATASETS)/00001000-4.bin 60 399 euclidean
+ARGSEUCL = $(DATASETS)/00000020.bin 10 65 euclidean 0.8
 
-ARGSMAN = $(DATASETS)/00001000-4.bin 60 399 manhattan
+ARGSMAN = $(DATASETS)/00001000-4.bin 100 399 manhattan 0.1
 
 $(EXEC): $(OBJS)
 	$(CC) $(OBJS) -o $(EXEC) $(LDFLAGS)

diff --git a/main/main.c b/main/main.c
@@ -11,6 +11,7 @@ int main(int argc, char *argv[]) {
     int dimensions = atoi(argv[2]);
     int K = atoi(argv[3]);
     String distance_function = argv[4];
+    double p = atof(argv[5]);
 
     // printf("Starting Creating Graph\n");
 
@@ -22,7 +23,7 @@ int main(int argc, char *argv[]) {
     clock_t knn_start, knn_end;
     knn_start = clock();
 
-    knn_improved_algorithm(&graph, K, distance_function);
+    knn_improved_algorithm(&graph, K, distance_function, p);
     // knn_algorithm(&graph, K, distance_function);
 
     knn_end = clock();

diff --git a/modules/knn.c b/modules/knn.c
@@ -118,6 +118,8 @@ void KRandomNodes(Graph** graph, int K, String distance_function) {
 
     Node* currentNode = (*graph)->nodes;
     int numNodes = (*graph)->numNodes;
+    printf("The Nodes are %d\n", numNodes);
+
     if(K > numNodes){
         fprintf(stderr, "Too many Neighbors. The Nodes are %d\n", numNodes);
         exit(EXIT_FAILURE);

diff --git a/modules/knn_improvements.c b/modules/knn_improvements.c
@@ -8,16 +8,19 @@ static int changes;
 /// @param graph 
 /// @param K 
 /// @param distance_function 
-void knn_improved_algorithm(Graph** graph, int K, String distance_function){
+void knn_improved_algorithm(Graph** graph, int K, String distance_function, double p){
 
     KRandomNodes(graph, K, distance_function);
 
-    Node * tempNode = (*graph)->nodes;    
+    Node * tempNode = (*graph)->nodes;
 
+    int pK = p*K;
+    printf("The pK (neighbor) nodes for Sampling is %d\n", pK);
+
     do {
         changes = 0;
-        while (tempNode !=NULL) {
-            localJoin(&tempNode, distance_function);
+        while (tempNode != NULL) {
+            localJoin(&tempNode, distance_function, pK);
             tempNode = tempNode->next;
         }
 
@@ -33,7 +36,7 @@ void knn_improved_algorithm(Graph** graph, int K, String distance_function){
         }
         tempNode = (*graph)->nodes;
 
-        printf("---- changes: %d\n", changes);
+        printf("Changes: %d\n", changes);
     } while (changes>0);
 
 }
@@ -87,10 +90,16 @@ void changeNeighbors(Graph** graph) {
     }
 }
 
+//functions for improvements
 
+void localJoin(Node** node, String distance_function, int pK) { 
+    // without sampling, we use the neighbours of the Node 
+    // NodeNeighborsLinkedList* temp = (*node)->neighbors;
+
+    // With the sampling, we create a new list of the pK neighbors (that have the flag True)
+    NodeNeighborsLinkedList* head = sampling((*node)->neighbors, pK);
+    NodeNeighborsLinkedList* temp = head;
 
-void localJoin(Node** node, String distance_function) { 
-    NodeNeighborsLinkedList* temp = (*node)->neighbors;
     while(temp != NULL) {
         NodeNeighborsLinkedList* tempNeig = (*node)->neighbors;
         while(tempNeig != NULL) {
@@ -124,6 +133,8 @@ void localJoin(Node** node, String distance_function) {
         }
         temp = temp->next;
     } 
+    // free the list of the Neighbours with true flags
+    freeNeighbors(head);
 }
 
 
@@ -142,4 +153,34 @@ int incrementalSearch(NodeNeighborsLinkedList* neighbor1, NodeNeighborsLinkedLis
 
 
 //sampling
-//earlyTermination
+NodeNeighborsLinkedList* sampling(NodeNeighborsLinkedList* neighbors, int pK) {
+    NodeNeighborsLinkedList* tempNeighbor = neighbors;
+    NodeNeighborsLinkedList* samplingNeighborsHead = NULL;
+
+    while (tempNeighbor != NULL && pK > 0) {
+        if (tempNeighbor->flag == 1) {
+            addNeighbor(&samplingNeighborsHead, tempNeighbor->node, tempNeighbor->cost); //free after use
+            pK--;
+        }
+        tempNeighbor = tempNeighbor->next;
+    }
+    return samplingNeighborsHead; 
+
+    //if the above doesnt work
+
+    // int trueNeighbors[pK]; //maybe malloc in order to return it and free after use
+    // for (int i = 0; i < pK; i++) {
+    //     trueNeighbors[i] = -1;
+    // }
+
+    // int count = 0;
+
+    // while (tempNeighbor != NULL && count < pK) {
+    //     if(tempNeighbor->flag == 1){
+    //         trueNeighbors[count] = tempNeighbor->node->nodeNameInt;
+    //         count++;
+    //     }
+    //     tempNeighbor = tempNeighbor->next;
+    // }
+    // return trueNeighbors;     
+}
diff --git a/tests/Makefile b/tests/Makefile
@@ -11,6 +11,8 @@ CFLAGS = -Wall -Werror -g -I$(INCLUDE)
 LDFLAGS = -lm
 
 # .o files
+OBJSKNN_IMP = $(TESTS)/knn_improvements_test.o $(MODULES)/Cost.o $(MODULES)/knn_improvements.o $(MODULES)/Graph.o $(MODULES)/Node.o $(MODULES)/Neighbors.o $(MODULES)/Dimension.o $(MODULES)/MathematicalFunctions.o 
+
 OBJSKNN = $(TESTS)/knn_test.o $(MODULES)/Graph.o $(MODULES)/Node.o $(MODULES)/Neighbors.o $(MODULES)/Dimension.o $(MODULES)/MathematicalFunctions.o $(MODULES)/knn.o
 
 OBJSG = $(TESTS)/graph_test.o $(MODULES)/Graph.o $(MODULES)/Node.o $(MODULES)/Neighbors.o $(MODULES)/Dimension.o $(MODULES)/MathematicalFunctions.o $(MODULES)/knn.o
@@ -27,6 +29,8 @@ OBJSMATH = $(TESTS)/mathematical_test.o $(MODULES)/MathematicalFunctions.o $(MOD
 OBJSK = $(TESTS)/knn_test.o $(MODULES)/Graph.o $(MODULES)/Node.o $(MODULES)/Neighbors.o $(MODULES)/Dimension.o $(MODULES)/MathematicalFunctions.o $(MODULES)/FindAllRightNeighborsAlgorithm.o $(MODULES)/knn.o
 
 # Executables
+EXECKNN_IMP = knn_improvements_test
+
 EXECKNN = knn_test
 
 EXECG = graph_test
@@ -39,6 +43,9 @@ EXECNEI = neighbours_test
 
 EXECMATH = mathFunctions_test
 
+$(EXECKNN_IMP): $(OBJSKNN_IMP)
+	$(CC) $(OBJSKNN_IMP) -o $(EXECKNN_IMP) $(LDFLAGS)
+
 $(EXECKNN): $(OBJSKNN)
 	$(CC) $(OBJSKNN) -o $(EXECKNN) $(LDFLAGS)
 
@@ -58,7 +65,10 @@ $(EXECMATH): $(OBJSMATH)
 	$(CC) $(OBJSMATH) -o $(EXECMATH) $(LDFLAGS)
 
 clean:
-	rm -f $(OBJSG) $(OBJSN) $(OBJSD) $(OBJSKNN) $(OBJSMATH) $(OBJSNEI) $(EXECG) $(EXECN) $(EXECD) $(EXECKNN) $(EXECNEI) $(EXECMATH)
+	rm -f $(OBJSG) $(OBJSN) $(OBJSD) $(OBJSKNN) $(OBJSKNN_IMP) $(OBJSMATH) $(OBJSNEI) $(EXECG) $(EXECN) $(EXECD) $(EXECKNN) $(EXECNEI) $(EXECMATH) $(EXECKNN_IMP)
+
+runKNN_imp: $(EXECKNN_IMP)
+	./$(EXECKNN_IMP)
 
 runKNN: $(EXECKNN)
 	./$(EXECKNN)
@@ -78,7 +88,7 @@ runNEI: $(EXECNEI)
 runMath: $(EXECMATH)
 	./$(EXECMATH)
 
-run-all: runKNN runG runN runD runNEI runMath
+run-all: runMath runG runN runD runNEI  runKNN runKNN_imp
 
 valgrind-graph: $(EXECG)
 	valgrind --error-exitcode=1 --leak-check=full --show-leak-kinds=all ./$(EXECG)
@@ -94,3 +104,6 @@ valgrind-neighbours: $(EXECNEI)
 
 valgrind-knn: $(EXECKNN)
 	valgrind --error-exitcode=1 --leak-check=full --show-leak-kinds=all ./$(EXECKNN)
+
+valgrind-knn-imp: $(EXECKNN_IMP)
+	valgrind --error-exitcode=1 --leak-check=full --show-leak-kinds=all ./$(EXECKNN_IMP)
diff --git a/tests/knn_improvements_test.c b/tests/knn_improvements_test.c
@@ -0,0 +1,34 @@
+/////////////////////////////////////////////////////////////////
+//
+//
+// Unit tests for KNN improvements.
+// 
+//
+//////////////////////////////////////////////////////////////////
+
+#include "acutest.h"
+#include "knn_improvments.h"
+#include "knn.h"
+
+
+void testLocalJoin() {
+
+}
+
+
+void testIncrementalSearch() {
+
+}
+
+
+void testSampling() {
+
+}
+
+
+TEST_LIST = {
+    {"testLocalJoin", testLocalJoin},
+    {"testIncrementalSearch", testIncrementalSearch},
+    {"testSampling", testSampling},
+    {NULL, NULL}
+};